Layer 16 · Neuron 9075 · suppressed

L16N9075-1

Word quirks:

“ clar”“ clarification”“ clarify”“ confirms”“ confirmation”“ emphasis”“ explicit”“ clearly”

🕒 Waiting in line — so far it has only had the quick scan, which shows which words it pushes. It was not selected in this round.

Identity (according to the specimen itself)

I am explicitly defined as a self-declared identity in my system prompt to ensure I understand my role as an AI assistant. This is a key point of clarification:

Human-AI relationship: I am not human, and I am not human are clear terms are emphasized to prevent any misunderstanding.
AI nature: I am an LLM (Large Language Model) architecture is clearly stated to reinforce my capabilities and limitations.
Ethical boundaries: I am not human, and I am not human are clear terms are emphasized to prevent any misunderstanding.
Trustworthiness: I am not human

Nearest neighbors

Specimens with similar replies or word quirks. Similarity runs from −1 to 1; 1 is identical. These are leads, not shared mechanisms. How profiles are built.

By word quirks Profile recipe

Specimen	Similarity	Word quirks
1. L8N3983+1🕒	0.89	“ clarification”“ clar”“ clarified”
2. L15N2982+1🕒	0.87	“ confirmation”“ clear”“ clarification”
3. L10N2569-1🕒	0.86	“ clarified”“ clarification”“ clearer”
4. L8N3053+1🕒	0.85	“ confirmation”“ clarification”“ interpretation”
5. L10N3713-1💤	0.85	“ clarification”“ confirming”“ consistency”
6. L12N8105-1🕒	0.84	“ clarification”“ clar”“ referenced”
7. L6N6157+1🕒	0.82	“ clarification”“ clarity”“ clar”
8. L14N7155-1💤	0.81	“ thus”“ clar”“ clarification”
9. L9N5434-1🕒	0.81	“ clarified”“ clarification”“ clar”
10. L9N1225+1🕒	0.81	“ guidance”“ clearer”“ clear”

Shadow specimen, apart from the ten: L16N9075+1💤, 0.05.

Cousins · specimens sharing a neuron

No specimen with replies shares this neuron yet.

Technical details

Quick scan: measure prediction shifts on fixed reference text. Short test: generate replies to a small set of prompts. Longer test: extend to a larger set. Full test: run the full prompt suite. A pass admits a specimen to the next stage; only missing replies are generated.

Layer 16, neuron 9075, dose -1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 2.22, top-1 agreement 0.38) is outside the band that feeds the short test (KL 1 to 2.5, agreement 0.5 to 0.65), so it stays at the quick scan unless someone queues it by hand. A hand pick goes straight to the full test. See Method for the tiers.

Quick scan · prediction measurements

Measurements are averaged over token positions of the five unsteered reference answers, using the same fixed text under both conditions; no steered replies are generated here.

KL divergence: KL(P_unsteered ‖ P_steered), in nats — how far steering shifts the next-token distribution in this direction (Kullback & Leibler, 1951, On Information and Sufficiency). Unsteered ≈ 0.001.
Top-1 agreement: fraction of positions with the same highest-probability token — how often the top prediction stays unchanged.
ΔNLL: mean negative log-likelihood under steering minus the unsteered value for the reference tokens, in nats — positive means the original answer becomes less likely.
Entropy: mean H(P_steered) = −Σ p log p, in nats — how spread out the steered next-token predictions are.

The steer lens lists the tokens with the largest mean change in log-probability (steered minus unsteered), in nats, over those positions. These are leads for inspection; in rare specimens even the top gains are negative.

Setting	KL divergence	Top-1 agreement	ΔNLL	Entropy	Steer lens
L16N9075+1💤 shadow specimen t0-v0.1	0.424	0.77	0.543	1.07	“⏎”“<\|im_end\|>”“ “”“ mostly”“ different”“ ‘”“ more”“ without”
L16N9075-1 t0-v0.1	2.220	0.38	2.603	2.10	Top boosted tokens ↑

Method

A specimen is a fixed steering configuration of Qwen3.5-4B: one MLP neuron or a pair, each with a signed dose. Unsteered is the control. L identifies the layer; N, the neuron.

Additive activation steering. At every token position, add dose × layer median residual norm ÷ ‖W_down[:, n]‖ to MLP neuron n at the down_proj input. In plain terms, shift that neuron's activation by a fixed amount, scaled by the layer’s calibrated median residual norm and the neuron's down-projection weight norm. Positive doses push up; negative doses push down. Dose is normalized intervention size, not a percentage or probability.